Note that some of the original submission proved redundant. For ease of presentation they have been removed. An explanation of what happened is set out here. (I have not changed the numbering.)
Introduction
My name is Brian Easton. I have a doctorate of science from the University of Canterbury and hold other qualifications in economics, mathematics and statistics. I hold honorary positions in six New Zealand universities including an adjunct chair at the Auckland University of Technology.
I was among the first to recognise that the largest group of New Zealand’s poor was children and their guardians/parents. My first published paper on poverty was more than forty years ago. Since then, when resources have allowed, I have extended my research on the measurement of poverty and inequality and its implications. My very long publications list on the topic of the bill can be supplied on request.
This submission is on my behalf but it also reflects the views and interests of those in the inequality research community
Summary
This submission broadly supports the Child Poverty Reduction Bill, setting it in a wider context.
It is in four parts.
Part I points out that the Child Poverty Reduction Bill is an extension of the Fiscal Responsibility Act (now incorporated into the Public Finance Act) but that it also represents a different vision of the role of the Government.
Part II explains that a poverty line is a measure of inequality and one which is consistent with earlier New Zealand social policy thinking.
Part III argues that the current measures of poverty in New Zealand can be improved and that it is imperative to do so if we want the reduction the bill promises to succeed.
Part IV discusses changes necessary to the Bill to implement the findings of the earlier part.
The submission concludes with a summary of this submission’s recommendations, including suggestions on how the bill should be amended.
The overall conclusion of this submission is that the Bill should be proceeded with, but that it should be amended as follows:.
- That all references to absolute poverty be eliminated from the Bill and that it be clear that the notion referred to in the Bill is relative poverty, perhaps as implicitly defined by the 1972 Royal Commission on Social Security.
- That the current processes of setting targets in the Bill be replaced as follows:
2a. That each incoming government be required to table in the House a set of targets for poverty reduction.
2b. That the government establish an independent technical committee to provide measures of the wellbeing of children and other New Zealanders.
2c. That the government establish an independent representative committee to use the measures provided by the technical committee to make recommendations on a suite of poverty lines which the government may use is setting its targets.
Part I: The Child Poverty Reduction Bill and the fiscal responsibility approach.
1.1. Both past New Zealand governments and some overseas ones have promised to eliminate child poverty or substantially reduce current levels. Necessarily this requires time but the target date mentioned in the promise has always been so far out that the politician or government making the promise will not be in power. Thus they have not been accountable for achieving the target and have not been held to account. Practically, such promises have been valueless.
1.2 This bill also makes such a promise, but the effect of the statute is that it holds every government between today’s and that at the target date accountable to Parliament in particular and the public in general. As such it has parallels to the 1991 Fiscal Responsibility Act (now incorporated into the Public Finance Act).
1.3 In 1994 the New Zealand Parliament enacted a Fiscal Responsibility Act, which required the government to set future targets for government debt. The government chooses the actual benchmarks but it is required to publish them. There is no legislated penalty for a failure to meet the debt targets – it is accepted that events outside the control of the government can impact on the outcome. Instead it is left to public opinion and parliament to assess the government’s success or otherwise in meeting its targets.
1.4 The parallels with the Child Poverty Reduction Bill are obvious. Again the government is required to set future targets but this time for numbers of children below a poverty level. Again the government chooses and publishes the benchmarks. There is no penalty for a failure to meet the targets for again, events outside the control of the government can impact on the outcome. Again, it is left for public opinion and parliament to assess the government’s success or otherwise in meeting its poverty reduction targets.
1.5 The Bill introduces a further explicit concern for fiscal policy. The 1994 Act was in the (neoliberal) framework of the government budget administered as if it were a household. This had represented a winding back from the earlier tradition that fiscal policy was concerned with wider social ends. (I can elaborate this if requested.) The effect of the Child Poverty Reduction Bill is to direct that fiscal policy must be concerned with wider social issues and not just ‘balancing the budget’. In particular it introduces inequality as an explicit policy objective.
1.6 Recently there has been concern among the public about the degree of inequality in New Zealand. In my opinion that concern has been sufficiently widespread for Parliament to be concerned with the issue. Not to be concerned would amount to its failing to represent the people of New Zealand.
1.7 The public concern about economic inequality has not been carefully analysed. As best as I can judge there are three major public threads (in addition to the personal belief that the individual is not getting an adequate share of total national income):
(i) There is a concern that the existing income (and wealth) distribution is not fair. This is essentially an ethical concern but also reflects a vision of the sort of society which New Zealand should be.
(ii) There is a concern that the existing income (and wealth) distribution threatens social coherence and hence social stability.
(iii) There is a concern that poverty among children (the largest group of the poor in New Zealand) affects their opportunity and life choices by compromising their health and educational attainment. This is not only detrimental to them (that is, it is not fair) but it results in poorer future economic performance (from the loss of human skills and from the higher social costs of dealing with the consequences of the earlier failure) and, ultimately, the long-term social coherence and viability of the nation. (I am reluctant to comment on the impact of poverty on criminal behaviour as the research evidence I know of is limited.)
1.8 Enacting the Child Poverty Reduction Bill would show Parliament taking up this challenge of inequality, especially in regard to the third concern mentioned in the previous paragraph.
1.9 Poverty and economic inequality are inextricably linked. The next part of this submission elaborates the linkage.
Part II: The relationship between measured poverty line and measured inequality.
2.1 Public discussion assumes that inequality is a well-defined notion which may be readily measured by a single indicator. (The Gini Coefficient is often cited, frequently by people who could not define it.)
2.2 The research is clear: there is no simple notion of inequality nor measure of it. This might be illustrated by the way in which on occasions various measures of inequality can be in conflict with each other, some suggesting it is increasing, some that it is decreasing and some showing no change. I can provide a lecture (or a course) on this topic, but perhaps a single illustration will suffice. In the mid-1990s some measures based on median incomes concluded that the transfer of income from the middle of the distribution to those with top incomes reduced poverty. (Median income fell, and therefore the poverty line base on median incomes fell so the numbers below the line fell.) This was despite an evident rise in hardship well attested by those working with the poor (for instance the number of food parcels distributed increased). In fact most other measures of poverty and inequality rose despite income inequality on most measures increasing. Casual use of measures of inequality and poverty can be treacherous.
2.3 However, if the shift in the income distribution is sufficiently large (as it was between 1985 and 1995), all sensible indicators of inequality will show a shift in the same direction. It is where the distribution shifts are small and short term that the indicators can be contradictory.
2.4 The niceties of research precision are not always helpful for those involved in policy implementation, even though they should be kept in mind. Practically, though, policy may have to choose a particular measure (or a suite of measures) of inequality and poverty. What the research says is that the choice requires a value judgement which cannot be objectively provided. However, research can help the subjective decision-makers to make an informed choice.
2.5 By proposing poverty lines, the Child Poverty Reduction Bill proposes a particular measure of inequality. If the select committee decides it wants to look at alternative measures, I am willing to assist it. However a poverty line has a logic derived from older social policy traditions which were abandoned in about 1990.
2.6 While the 1972 Royal Commission on Social Security was nominally concerned with the social security system, it actually presented perhaps the most comprehensive view of the economic foundations of social policy of any official New Zealand body. In particular it said (original’s underlining)
The aims of the system, should be
(i) First, to enable everyone to sustain life and health;
(ii) Second, to ensure, within limitations which may be imposed by physical or other disabilities, that everyone is able to enjoy a standard of living much like that of the rest of the community, and thus is able to feel a sense of participation in and belonging to the community.
(iii) Third, where income maintenance alone is insufficient (for example, for a physically disabled person), to improve by other means, and as far as possible, the quality of life available.
(2.7 The Royal Commission used the term ‘participation in and belonging to’. The second part of the expression has sometimes been omitted. The reason for doing so is not always made clear but a possible one is that it is redundant – that participation in a community means one belongs to it. In deference to the original formulation this paper uses the full expression.)
2.8 The first aim of the Royal Commission is consistent with a notion of maintaining everybody at or above an absolute level of poverty; its second aim defines a notion of relative poverty. The rest of this submission has much to say about this distinction. The Royal Commission saw the second notion as the basis of New Zealand social policy.
(2.9 We will not pursue in detail here the third aim although it is very relevant for social policy. For instance, no amount of income will resolve poor access to a building for those with mobility challenges.)
2.10 Instructively, the Royal Commission did not say anything about the top of the income distribution. Perhaps it was because that lay outside the terms of reference of the enquiry but there may have been a deeper reason.
2.11 In 1971, about the time that the Royal Commission was completing its report, American philosopher John Rawls published his seminal Theory of Justice. He argued for ‘justice as fairness’, recommending equal basic rights, equality of opportunity, and the promotion of the interests of the least advantaged members of society. In particular he proposed a ‘maximin’ principle – that the system should be designed to maximize the position of those who will be worst off in it – which he derived by individuals choosing to improve the position of the worst off in society with the possibility they might find themselves in that position.
2.12 There is no evidence that the Royal Commission had access to Rawls’ thinking but there is an uncanny parallel between its conclusions and those of The Theory of Justice. Had the book’s arguments been presented to it, it seems likely that the Royal Commission would have comfortably and seamlessly incorporated the Rawlsian principles into its report.
2.13 In particular, the Royal Commission’s second aim of the system – that everyone should be able to enjoy a standard of living much like that of the rest of the community, and thus able to feel a sense of participation in and belonging to the community – is a reformulation of the ‘maximin’ principle. Note that Rawls does not say much about the top of the income distribution. His focus is on the providing an income floor.
2.14 So it is completely in keeping with New Zealand’s traditional social policy objectives to focus on providing a minimum income floor – in effect a poverty line – without too much attention to further up the distribution (other than as funders of the floor).
2.15 Thus the focus of the Child Poverty Reduction Bill on a poverty line is consistent with both New Zealand’s traditional approach to social policy and reputable philosophical thinking.
2.16 Given the size of child poverty – the exact numbers of those defined being in poverty depend on the choice of the poverty line (to the numbers should be added their guardians/parents) – I am confident that a significant reduction in the numbers below any sensible poverty line will show a reduction in overall income inequality on any other sensible measure.
(2.17 While well over half of the poor are children and their guardians/parents there remain others who do not belong to this category and are unlikely to be affected by any policy changes which reduce child poverty. Should their needs be addressed? For many of them – including the sick and the disabled – the answer is ‘yes’; it is to be hoped that their needs will be considered in due course. However, the RCSS/Rawls approach does not place an obligation to ensure that everyone below the poverty line should receive more income assistance. Aside from those for whom employment is a more appropriate solution, there are also those who choose low income life styles as the best way they can participate in and belong to their community.)
Part III: Setting a Poverty Line.
3.1 The way that the numbers in poverty are usually counted is as follows. An indicator of living standards is identified and measured for each household.[1] Most often this has been disposable income – market income plus benefit income less direct taxes paid – which will be used here to illustrate the measurement principles. Then the aggregate indicator (disposable income) is adjusted for household composition. (It is easiest to think of this in the case of being converted to per capita income but in practice children are weighted less than adults and there is an allowance for economies of scale.) The resulting aggregate ‘equivalised’ income is attributed to each member of the household (this assumes that the income/expenditure of the household is shared equally – it may not be). A poverty line which is an income level is chosen and the number of people (or children or whatever) living in households below the poverty level is the poverty count.
3.2 To summarise schematically for a household:
Market Income
plus
Benefit Income
less
Direct Taxes paid
equals
Household Disposable Income
adjust for
household composition (typically using ‘equivalence scales’)
equals
Equivalised Household Disposable Income (EHDI)
choose
a poverty line
count
the numbers in households whose EHDI is below poverty line
which gives
the poverty count
3.3 The method was devised in the 1970s based upon research overseas. It has hardly changed since, despite a number of obvious weaknesses. Most were known in the 1970s, but the data available was limited so that it was not possible to address the problems then. What is disappointing is that there has been hardly any improvement in the subsequent 40 years even as better data comes available. Here are some examples of the weakness of the method.
3.4 Is disposable income the right measure? For instance, if the government was to eliminate school fees it would make no difference to the poverty count, despite families being under less financial stress. On the other hand, if the subsidies to primary health care were withdrawn it would place families under greater stress but, again it would make no change to the poverty count. Childcare costs are not deducted from income; a family may be impoverished by them as a consequence of earning income, but it is treated in poverty count terms exactly the same as family which gets free childcare or does not incur such costs. The costs of housing raises a series of problems – contrast the family in their own mortgage-free home with one paying an onerous amount for their rent. (Deducting housing costs from income is obviously wrong, since it is subtracting an expenditure from income – apples from oranges.)
3.5 The equivalence scales used to adjust for housing composition are problematic. For instance it would appear that the household economies of scale which are typically assumed (the equivalence scales usually used are not estimated but invented) overestimate the effective income of large households and underestimate it for small households, affecting the composition of the poor (but probably not the poverty count by much).
3.6 These measurement failures in no way affect the actual poverty. They mean that our estimates are distorted and it seems likely that any consequent measures to reduce poverty are less effective – more expensive – than they need be.
3.7 The choice of poverty line is also deeply problematic when, as typically occurs, it is based on opinion with little empirical underpinning. The original level used was the one implicitly proposed by the 1972 Royal Commission on Social Security which was presented with considerable evidence of family economic circumstances. As emphasised earlier, choosing a poverty line involves judgment but, compared to today’s proposed levels, the Royal Commission’s was informed by actual evidence.
3.8 Given the unsatisfactory theoretical and empirical bases of the available poverty lines, including those mentioned in the Bill (Part 2), it may be wondered whether they are of any use or when policy weight should be placed upon them. The short answer is that they are the best we have got, and will have to do in the interim until we have a better one.
3.9 In order to overcome some of the weaknesses of the current measures and to enable easier comparisons as better measures are introduced, the published targets should be in terms of proportional reductions (e.g. half on the chosen measure by the given date) rather than a reduction in numbers.
3.10 Note that poor quality poverty measures will lead to inefficient targeting, especially of the wrong groups. An even greater fear is that they can be abused for policy purposes. For instance, reducing or restraining those on middle incomes reduces poverty according to measures based on the median income without necessarily having any effect on the living circumstances of the poor (as unintentionally occurred in the early 1990s). Raising the costs of school fees or health care will not effect the measured number of the poor although they will be in greater hardship. Changes in housing policy and circumstances are likely to have an erratic impact on the measures even if they reduce hardship. While this may not be the intention of this government, the legislation needs to ensure a less benign government cannot manipulate the faulty measures to avoid addressing actual poverty.
3.11 How might we construct and implement the monitoring of better poverty measures in the future?
Part IV: Setting a Better Poverty Line.
4.1 It is indicative of how superficial are current poverty lines that most commentators do not link the measure to any fundamental notion of poverty. The bill does attempt to do this.
4.2 But it does so by focusing on material hardship. This, in fact, is a notion of absolute poverty. For as the country prospers, a material hardship-based poverty line need not be adjusted. We would be consolidating the 1990 ‘Redesigning the Welfare State’ approach which has been to index benefit levels to consumer inflation only.
4.3 However the proposed poverty measures seem to be based on the notion of relative poverty given that they imply that as prosperity rises so does the poverty line. I leave others to reconcile the contradictions.
4.4 The following explains how a less opinionated, more evidence-based poverty level might be derived. Without resiling from the earlier point that there has to be a value judgement, it suggests how the judgement can be considerably more informed than is the current practice.
4.5 This submission takes as its fundamental notion of poverty that set out by the 1972 Royal Commission on Social Security of those in (relative) poverty having an inadequate standard of living unlike that of the rest of the community to the extent that they are unable to feel a sense of participation in or belong to the community.
4.6 Observe that such poverty compromises the opportunity for the child to develop as set out by the Fraser-Beeby principle (updated, as Clarence Beeby would wish, to be gender neutral with the key notion underlined):
The government’s objective, broadly expressed, is that every person, whatever her or his level of academic ability, whether he or she be rich or poor, whether he or she live in country or town, has a right, as a citizen, to a free education of the kind for which he or she is best fitted and to the fullest extent of her or his power.
4.7 It would be possible to ask people to what extent they felt they were able to participate in and belong to their community. As far as I know, that has never been done systematically and, in any case, the response would be subjective and consequently subject to all the problems of policy using subjective measures.
4.8 Instead there are objective measures of standards of living. In particular New Zealand has a Living Standards Research program. The 2016 Household Economic Survey asked about 20 child-specific items as follows:
Does the household or child have
two pairs good shoes for each child
two sets of warm winter clothes for each child
waterproof coat for each child
all the uniform required by the schools
a separate bed for each child
fresh fruit and vegetables daily
a meal with meat, fish or chicken (or vegetarian equivalent) at least each second day
a range of books at home suitable for their ages
a suitable place at home to do school homework
friends around to play and eat from time to time
friends around for a birthday party
good access at home to a computer and internet for homework
mobile phone if aged 11+?
In order to keep down costs to help in paying for (other) basic items (not just to be thrifty or to save for a trip or other non-essential) has the household taken the following economising actions:
postponed visits to doctor
postponed visits to dentist
unable to pay for school trips/events for each child
had to limit children’s involvement in sport
children had to go without music, dance, kapa haka, art, swimming or other special interest lessons
children continued wearing worn out/wrong size clothes and shoes
made do with very limited space to study or play?
4.9 These items could be used to construct a wellbeing index for children. The index could be compared with the household income to help form a poverty line.
4.10 For instance, suppose it was decided that without at least two pairs of good shoes a child would have considerable difficulties participating in and belonging to their community and it was found that standard households (say two adults and two children) which had disposable incomes below $X per week lacked two pairs of shoes for each child. There follows a strong case for setting the poverty line at $X per week (or higher) for this household type. (Shoes here are the first item on the list and used only as an illustration. Historically, researchers used as an indicator when families did not take their children to their doctor because of financial hardship. This ‘economising action’ was a last or desperate resort. Such thinking has been a major factor behind the policy of higher subsidies for children visiting their GP. )
4.11 The exercise is more complicated than this, because all 20 items should be used (at least to some extent).[2] But in principle what can be done is that a tabulation could be constructed which gave the probability of a household being below each income level. An appropriate panel could then identify an income level below which they judged that the household was so deprived that the children were suffering poverty, thereby setting a poverty line. Observe that there is still a judgement to be made but that it would be an informed judgement and not just uninformed opinion.
4.12 This approach requires a major change to the monitoring and report sections in the Bill. First, it is necessary to define poverty more carefully. The bill defines it as ‘material hardship’ but there are other hardships. For example, being unable to invite friends around for a birthday party – reciprocating invitations to friends’ parties – is not material hardship, it is a humiliation. It can also be development threatening. If a schoolgirl cannot go with her friends to, say, a significant gig, it not only makes her an outsider in the class but is likely to turn her away from that community and undermine her educational performance. (That is why teachers have been known to provide the means to participate with the group for a promising pupil from their own pocket.)
4.13 The poverty targets need to be articulate in relationship to a coherent definition of poverty. The above illustrates how that could be done with a particular definition, but the method and process is quite general and could be applied for any other definition (providing the data was available).
4.14 While the current Bill makes no provision for a technical group or an advisory group to assist with the judgement this should be provided for in the revised Bill. Setting poverty targets requires greater technical skills than are currently in the public service (although there are some public servants whose work in the area is to be greatly admired).
4.15 Public servants are not well paced to advise directly on judging the poverty line without the intermediation of a consultative committee. In particular it puts the Government Statistician into an invidious position of making judgements which are not technical. In order to maintain the integrity of the statistical base the Government Statistician should be shielded from controversial decisions. The proper function of the Government Statistician in this context is as an adviser to ministers and the executive. It is a conflict of interest to have her or him also advising Parliament which is concerned with keeping the government accountable.
4.16 In any case the poverty targets should be shifted out of the main Bill to a schedule to the Bill. The government should have the option of revising the schedule, tabling the revision to Parliament for debate, as better measures become available. There should also be a requirement that an incoming government should submit its own (revised) schedule within six months of taking office.
Summary
6.1 The proposed Child Poverty Reduction Bill enables the people of New Zealand, especially through its representatives in Parliament, to hold the government to account in regard to two important issues:
(i) the level of economic inequality in New Zealand;
(ii) the hardship among children and their families not only effects their wellbeing but compromises the long-term development of New Zealand.
It should be proceeded with following the amendments set below.
6.2 The Bill adds a further responsibility to fiscal policy in addition to the (now merged) 1994 Fiscal Responsibility Act by explicitly adding a dimension of social wellbeing to narrower budget concerns. As such, it formally broadens the notion of fiscal responsibility.
6.3 Setting a target for reducing poverty among New Zealand’s children makes sense because:
(i) given its size and extent, child poverty is almost certainly the most serious issue of wellbeing in New Zealand;
(ii) reducing child poverty will have long-term benefits to New Zealand including improving the nation’s economic and social performance and its sustainability;
(iii) substantially reducing child poverty will also substantially reduce economic inequality in New Zealand.
Even so, the needs of smaller groups who are also in poverty through no fault of their own – such as from disability and sickness – should not be ignored.
6.4 The Bill is currently ambiguous as to whether it is concerned with absolute poverty – that is, individuals are in material hardship – or relative poverty, that is, whether they have sufficient to participate in and belong to their community and able to share in its progress. The Bill should make it clear that it is the second notion with which it is concerned.
6.5 The current measures of poverty levels and the numbers they report are imperfect. While they are useful for public discussion they are not robust for public policy or research. As such they need to be improved.
6.6 In the interim the flawed measures will have to be used for targeting purposes. It makes sense to choose a suite of targets, not only to reduce the inadequacy of each measure but to limit the ability of any government to manipulate policies to attain the particularities of the individual targets while ignoring the spirit of the exercise to improve the wellbeing of children. .
6.7 In order to replace the flawed measures with higher-quality ones the government should establish a working group of technically competent researchers. They should prepare improved measures (although they will be limited by data availability). However, the research working group should not make recommendations on the precise poverty line.
6.8 Instead, there should be established an advisory group to assess the research findings and recommend to Parliament and the government a suite of robust poverty lines. The advisory group should be representative by such social characteristics as age, ethnicity, family experience and gender. (However the chair of the research working group may be appointed to it in order give the lay group better access to the research findings.) Given a commitment, the work program and the decisions which evolve out of it should enable the incoming government in 2020 to set out revised poverty targets.
6.9 The import of these recommendations is that Parliament should not only pass the Child Poverty Reduction Bill, albeit with some improvements in the evaluation, monitoring and reporting, but it should also ensure that the measures necessary to implement it effectively are taken (including ensuring that there is enough funding to enable the research working group to work quickly and efficiently).
Recommend Changes to the Bill
7..1 The drafting of statutes involves specialist skills which I do not have. The following are intended to guide the those drafting the revised bill..
7..2 That all references to absolute poverty be eliminated from the Bill and that it be clear that notion is relative poverty perhaps as implicitly defined by the 1972 Royal Commission on Social Security.
7..3 That the current processes of setting targets in the Bill be replaced as follows;
7..3a That each incoming government be required to table in Parliament a set of targets for poverty reduction.
7..3b That the Government establish an independent technical committee to provide measures of the wellbeing of children and other New Zealanders.
7..3c That the Government establish an independent representative committee to use the measures provided by the technical committee to make recommendations on a suite of poverty lines which the Government may use when setting its targets. .
Endnotes
[1] Typically, the data source is the Household Economic Survey. This or other surveys are smallish samples and so the results are subject to error. Another source of error is that the income variables, in particular, are usually self-reported. And may not be very accurate.
[2] The construction of high-quality statistically robust indexes is a technical challenge. Currently all that is available are rather primitive versions. This is because of an unwillingness to provide the resources to do the analysis; New Zealand certainly has the statistical expertise to do better if the resources were available.